Why ReLU networks yield high-confidence predictions far away from the training data and how to mitigate the problem

dc.contributor.author	Hein, Matthias
dc.contributor.author	Bitterwolf, Julian
dc.date.accessioned	2020-10-02T15:11:02Z
dc.date.available	2020-10-02T15:11:02Z
dc.date.issued	2019
dc.identifier.isbn	978-1-7281-3293-8
dc.identifier.uri	http://hdl.handle.net/10900/107479
dc.language.iso	en	de_DE
dc.relation.uri	http://dx.doi.org/10.1109/CVPR.2019.00013	de_DE
dc.subject.ddc	004	de_DE
dc.title	Why ReLU networks yield high-confidence predictions far away from the training data and how to mitigate the problem	de_DE
dc.type	Article	de_DE
dc.type	ConferenceObject	de_DE
utue.publikation.seiten	41-50	de_DE
utue.personen.roh	Hein, Matthias
utue.personen.roh	Andriushchenko, Maksym
utue.personen.roh	Bitterwolf, Julian
dcterms.isPartOf.ZSTitelID	2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)	de_DE